PRISM: PSI and Voronoi diagram based Automated Exposure Notification with location privacy

Abstract

Automated Exposure Notification techniques have been developed as pervasive risk assessment tools in public crisis management, particularly for the containment of infectious diseases. However, the widely adopted Bluetooth Low Energy or location-based AEN techniques face challenges in simultaneously providing fined-grained, real-time, and location privacy-preserving detection and notification. Therefore, we propose PRISM, an enhanced and hierarchical AEN scheme, offering both efficiency and security. Specifically, by analyzing the geographic properties of contact tracing in crisis, we first construct a Voronoi diagram map for detailed disaster-related data collection and management. Then, we introduce Hazardous Monitoring for broadcasting disaster information, which enables citizens to conveniently receive updates from public crisis management departments and rapidly recognize potential risks. In addition, we design Secure Exposure Identification, leveraging the Voronoi diagram and homomorphic encryption-based Private Set Intersection (PSI), allowing citizens to securely trace and identify contact with infected individuals. Detailed security analysis confirms that PRISM effectively safeguards citizens against privacy breaches. Finally, a PRISM system is developed and implemented for testing under real-world constraints, with experiment results indicating its efficacy and operational efficiency.